The discovery of tumour markers has revolutionised the detection, diagnosis and management of some cancers.
As with any test, there are caveats, and a poor understanding of these may lead to misuse of the test, false reassurance and consequent underdiagnosis of cancer, or false positives and overinvestigation of patients.
A recent hospital audit concluded that 90 per cent of requests for tumour markers were inappropriate, resulting in about £250,000 of wasted healthcare expenditure per year.1 Given the pressure on GPs to refer less, prescribe less and request fewer investigations, the need to use tests appropriately has never been greater.
1. Uses of tumour markers
Most tumour markers are proteins, often produced by tumour cells. They may serve to aid diagnosis, stage disease or augment tumour monitoring and response to treatment.
In their book on cancer and its management, Souhami and Tobias described the ideal tumour marker.2 Reality, however, rarely reflects the ideal and tumour markers are no exception. In an ideal world, each tumour marker would be 100 per cent specific and sensitive to the tumour in question, with a 100 per cent negative and 100 per cent positive predictive value, and levels that reflect tumour mass and recurrence of disease; the tumour would also be treatable.
In addition, the test should be cost-effective, acceptable for the patient, repeatable and appropriate for screening purposes.
One of the most commonly requested tumour markers is PSA. As its name suggests, it is organ specific, but it is also commonly raised in benign conditions of the prostate such as BPH, UTI, acute urinary retention, prostatitis and after urethral catheterisation.
The National Academy of Clinical Biochemistry (NACB) recommends that PSA should not be used for screening purposes and its use as a screening tool remains controversial in the medical literature.
PSA testing is, however, helpful in diagnosis if used in the context of clinical suspicion. It is also used in the detection of recurrence, monitoring and predicting prognosis.1
PSA may be falsely low by up to 50 per cent in patients taking 5-alpha reductase inhibitors such as finasteride. There is also evidence to suggest that concurrent statin use may lower PSA levels and that statins should be stopped in the weeks before testing.
Although higher levels of PSA increase the likelihood of malignancy, about 15 per cent of men with histological evidence of prostate cancer have a PSA <4g/L. This may be partly explained by highly undifferentiated disease.
Despite this, 25 per cent of men with a PSA of 4-10g/L and 50 per cent of men with a PSA >10g/L will have malignancy. PSA levels in excess of 100g/L are usually, but not exclusively, suggestive of metastases.
3. Carcinoembryonic antigen (CEA)
Doctors have been criticised for requesting tumour markers in patients without symptoms suggestive of a specific cancer and for inappropriately using markers to diagnose cancers.
CEA is often misused in this way. It should not be routinely used in screening or diagnosis and its uses mainly lie in establishing prognosis, monitoring disease and assessing response to treatment.
A raised CEA does not always suggest malignancy and normal levels do not exclude sinister disease.
CEA is higher in smokers and apart from colorectal cancer, cancer of the ovary, stomach, oesophagus, pancreas, lung, mesothelium and breast may also increase CEA levels, particularly in advanced disease.
Non-malignant causes of a raised CEA include ulcerative colitis, gastritis and pancreatitis.
The Dukes staging classification is now gradually being replaced by the tumour, node, metastasis (TNM) classification.
Reliance on CEA to exclude colorectal cancers could miss a significant number of cases, even in those patients with metastatic disease. In addition, CEA can be raised more readily in right rather than left-sided colonic tumours.
4. Cancer antigen (CA) 125
The use of CA125 as a screening tool for ovarian malignancy is under review and data will not be published until 2015.
The NACB, however, advocates its use in diagnosis as well as for prognosis and monitoring. If there is clinical suspicion of malignancy, a pelvic ultrasound scan should be organised in addition to CA125 levels.
It should be borne in mind that CA125 is also raised in a number of other cancers, including endometrial, cervical, breast, hepatocellular, pulmonary, pancreatic and non-Hodgkin's lymphoma (see box below).
|Uses of tumour markers|
|Tumour marker||Relevant cancer||Other cancers in which marker may be raised|
|Alpha-fetoprotein||Germ cell tumour/hepatocellular carcinoma||Colorectal, gastric, lung, pulmonary|
|Calcitonin||Medullary thyroid carcinoma||None known|
|Cancer antigen 125 (CA125)||Ovarian cancer||Breast, cervical, endometrial, hepatocellular, lung,
non-Hodgkin’s lymphoma, pancreatic, peritoneal, uterus
|Cancer antigen 15-3 (CA15-3)||Breast cancer||None known|
|Cancer antigen 19-9 (CA19-9)||Pancreatic cancer||Colorectal, gastric, hepatocellular, oesophageal, ovarian|
|Carcinoembryonic antigen (CEA)||Colorectal cancer||Breast, gastric, lung, mesothelioma, oesophageal, pancreatic|
|hCG||Germ cell and testicular carcinomas/gestational trophoblastic neoplasia||Lung|
|Paraproteins (M protein,
Bence-Jones protein); also measured in urine
|B cell proliferative disorders (such as multiple myeloma)||None known|
|PSA||Prostate cancer||None known|
|Thyroglobulin||Thyroid cancer||None known|
Non-malignant causes of a raised CA125 include TB, viral hepatitis, pancreatitis, acute urinary retention, chronic renal failure, colitis, cardiac failure, pericarditis, cystic fibrosis, diverticulitis, endometriosis, irritable bowel syndrome, fibroids, ovarian hyperstimulation, menstruation, pregnancy, pneumonia, sarcoid and lupus.
A full discussion of tumour markers is beyond the scope of this article. It is clear, however, that the decision to measure a tumour marker must be made on an individual basis, with consideration of the clinical picture and how the test will influence management, and an understanding of any pitfalls.
Indiscriminate testing with a panel of tumour markers is unhelpful.
- Dr Thakkar is a GP in Wooburn Green, Buckinghamshire
1. Sturgeon CM, Lai LC, Duffy MJ. Serum tumour markers: how to order and interpret them. BMJ 2009; 339: b3527
2. Souhami R, Tobias J. Cancer and its management (fifth edition). Oxford, Wiley-Blackwell, 2005